Corrosion rate estimating device and method

Patent metadata

Field	Value
Publication number	US-11892392-B2
Application number	US-201917058098-A
Country	US
Kind code	B2
Filing date	May 24, 2019
Priority date	May 24, 2018
Publication date	Feb 6, 2024
Grant date	Feb 6, 2024

Abstract

A corrosion rate estimation apparatus includes a corrosion rate measurement unit adapted to repeat water supply cycles in which water is supplied to soil in which metal to be evaluated is buried and measure a cycle number of each water supply cycle, one or more time points in each cycle, and a corrosion rate of the metal at the time point(s); a learning unit adapted to find a prediction model by accepting the cycle number, the time point(s) in the cycle, and the corrosion rate(s) as input and using a machine learning algorithm, the prediction model representing a future corrosion rate; and a corrosion rate estimation unit adapted to assign a cycle number and a time point in the cycle at which a corrosion rate is desired to be estimated to the prediction model and estimate a corrosion rate of the metal at the time point.

First claim

The invention claimed is:

1. A corrosion rate estimation apparatus comprising:
a corrosion rate measurement unit adapted to repeat water supply cycles in which water is supplied to soil in which a metal to be evaluated is buried, and to measure a cycle number of each water supply cycle, one or more time points in each water supply cycle, and a corrosion rate of the metal at the one or more time points;
a learning unit adapted to find a prediction model using the cycle number, the one or more time points, and the corrosion rate as input to a machine learning algorithm, the prediction model representing a future corrosion rate; and
a corrosion rate estimation unit adapted to assign a particular cycle number and a time point in a particular cycle at which a desired corrosion rate is to be estimated to the prediction model, and to estimate the desired corrosion rate of the metal at the time point.

2. The corrosion rate estimation apparatus according to claim 1 , further comprising a variation function calculation unit adapted to calculate a variation function that represents changes in a plurality of corrosion rates measured by the corrosion rate measurement unit, and to provide the variation function to the learning unit, wherein
the learning unit accepts the variation function as input and finds the prediction model representing the future corrosion rate using the machine learning algorithm.

3. A corrosion rate estimation method performed by a corrosion rate estimation apparatus, the corrosion rate estimation method comprising:
repeating water supply cycles in which water is supplied to soil in which metal to be evaluated is buried and measuring a corrosion rate of the metal in each water supply cycle a plurality of times;
finding a prediction model by accepting as inputs to a machine learning algorithm, cycle numbers of the water supply cycles and corrosion rates, the prediction model representing a future corrosion rate; and
assigning a cycle number and a time point in a cycle at which a desired corrosion rate is to be estimated to the prediction model, and estimating the desired corrosion rate of the metal at the time point.

Assignees

• Nippon Telegraph & Telephone

Inventors

• Mineta Shingo
• Oki Shota
• Mizunuma Mamoru
• Oka Soichi
• Sawada Takashi

Classifications

• G01N17/04 Primary

  Corrosion probes · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G06N20/00

  Machine learning · CPC title

• G01N17/006 Primary

  of metals · CPC title

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title